Citation

Stouffer SA. Am. J. Sociol. 1933; 39(2): 180-193.

Copyright

(Copyright © 1933, University of Chicago Press)

DOI

10.1086/216369

PMID

unavailable

Abstract

For statistical treatment of sociological data when the dependent variable is on a quantitative scale and the independent variables are not, it is first important to test the significance of the difference between means in sub-classes. By new and simple procedures the significance of the sums of differences between averages and of differences between differences, holding constant other factors, may now be tested. Application of these procedures to test scores of attitudes toward prohibition of 238 students of the University of Chicago in 1929 shows that, taken as a whole, the men were wetter than the women; students whose fathers were not reported as voting dry were wetter than students whose fathers were reported as voting dry; and students from neighborhoods not reported as dry were wetter than students from neighborhoods reported as dry. In each case the other two factors were held constant. Application of the same procedures to case-history rating led to nearly identical conclusions, indicating that as far as results were concerned it made little difference whether test or case-history methods were used.